Interference can be a major impediment to the operational performance of radar systems. A common source of interference is impulsive noise which may appear due to regional lighting discharges or local man-made noise sources, for example. Impulsive noise is characterized as having a short duration and therefore only affects a small portion of a time series of pulses. However, for systems that integrate signals detected over a long period together, such a short duration of interference can result in elevated noise for an entire integration period. Traditional methods for mitigating signal degradation due to impulsive noise relies on the detection and removal of corrupted pulses and the use of various techniques to limit frequency leakage over the signal band. These techniques typically rely upon time domain envelope detection methods within the bandwidth of operation. For example, such a technique may involve analyzing the received envelope for a signal larger than expected in order to detect interference within the time segment.
However significant degradation may occur even with impulse events that do not significantly increase the envelop power of the received signal (see, e.g., FIG. 6A). The received signal may have significant energy from radar clutter and other sources, for example, and thus rendering these envelop detection techniques ineffective for lower amplitude events. Moreover, the effective signal to noise ratio may also be adversely affected by these low amplitude events.
For a high frequency radar operating with coherent integration times of a several minutes, for example, impulsive noise can corrupt significant portions of the received time sequence. In Doppler radar, data is converted from the time domain to the frequency domain by Fourier transformation. This has the effect of smearing the impulsive energy in the spectral domain thus decreasing the probability of target detection at all Doppler frequencies.
Detecting Impulses in the time domain is problematic and limits excision to larger spikes that are generally the result of local lightning storms. However, even small impulses can significantly elevate the apparent noise level and hence degrade the signal-to-noise level. Detection of these small impulses is best accomplished before narrow-band filtering.
Thus, improved methods of detecting and mitigating impulsive noise events are desired.